SDOH. It’s the acronym for one of the hottest topics in healthcare today, and it should be. In fact, multiple studies show that social determinants of health (SDOH) significantly affect a person’s well-being.
Stakeholders across the healthcare realm—government agencies, payers, providers, data scientists, health researchers—are looking to identify the impact of SDOH on clinical outcomes and cost of care. Most importantly, all stakeholders have a responsibility to identify their patients’ social vulnerabilities. In doing so, they should collaborate to improve the quality of life for the people they serve—especially those with complex, costly care needs.
Today, the healthcare industry is increasingly turning to artificial intelligence and machine learning to make sense of oceans of SDOH data, to make informed decisions and take effective action. At HVH, we’ve modernized the approach to data analytics and machine learning for SDOH. Using activity-based intelligence, our platform identifies high-cost, high-need (HCHN) patients, to optimize and guide the design of interventions that will help patients most effectively.
Nearly 80 percent of health outcomes are influenced by nonclinical factors1
The World Health Organization defines social determinants of health as the conditions in which people are born, grow, live, work, and age. These circumstances are shaped by the distribution of money, power, and resources at global, national, and local levels. In fact, we know that clinical health is responsible for only 10 to 15 percent of a person’s well-being. Factors other than medical care—genetics, social, environmental, behavioral—are known to play a more significant role in population and individual health. Fundamentally, people who face unmet social and financial needs have a different type of healthcare experience than those who do not face such challenges.
Among the unmet social needs that impact the utilization of healthcare services are food environment, community safety, housing, social support, and transportation. For example, people who face public safety issues are three times more likely to have multiple visits to the emergency room (ER); and those with food or transportation challenges are nearly 2.5 times more likely to visit the ER multiple times.
Evidence shows that people who use the ER and hospital more often than the general population, are those who live with three or more chronic diseases and functional limitations. In addition, the median household income among these patients is less than half that of the average American household. Therefore, financial instability, along with unmet social needs, leads to increased use of healthcare services.
Activity-based intelligence: Finding and supporting HCHN patients
Until now, insurance companies and clinical researchers have relied on traditional data analytics methodologies to manage patients with a high-cost, high-need profile. Traditional analytics are based on statistics using old patterns and information. Most of today’s predictive models are linear in design and have limitations because of variables with limited predictive accuracy. Namely, they focus on specific populations and specific interventions and lack predictive capabilities at the patient-level. In comparison, current thinking supports that predictive analytics should gather information from multiple sources—clinical, claims, consumer, socioeconomics, media, environmental.
HVH’s powerful AI/machine learning platform uses a modern analysis methodology known as activity-based intelligence. Our platform rapidly integrates data from multiple sources to discover relevant patterns. These patterns then provide insight into what is predictable and actionable. In real time, our predictive models identify, with high fidelity, patients who are most likely to benefit from redesigned care delivery systems and care management.
Specifically, our methodology facilitates health enterprise zone initiatives to improve access to health care and health outcomes in underserved communities by designing detailed heat maps that identify unmet social needs and neighborhood stress concentrations. Our models optimize and guide the design of measurable interventions that are effective in improving social, personal, and health conditions.
1Aroditis A. Leveraging Social Determinants of Health Data to Improve Accountable Care Delivery and Gain a Complete Picture of Patients’ Needs #NHITWeek. HIMSS. Published November 12, 2018.